# Building models and plotting the results
def plot_comparison(area_tessellation, n_agents, show_real):
if area_tessellation in dict_area_tess_names.keys():
real_measures, sepr_measures, depr_measures, ditras_measures = build_model(area_tessellation, n_agents, n_individuals = number_individuals)
else:
return print("Please input a valid area and tessellation pair, e.g. 'a1_t1' to work on NY State with squared tessellation")
list_measures = ["Travel Distance (Jump Length)",
"Radius of Gyration",
"Uncorrelated Entropy",
"Distinct visited Locations",
"Visits per Location",
"Location Frequency"]
list_labels = ["Δr", "r_g", "S^{unc}", "N_u", "V_l", "L_i"]
dict_models_measures = {"real": real_measures, "sepr": sepr_measures, "depr": depr_measures, "ditras": ditras_measures}
# Getting x and y for each distribution
dict_x, dict_y = get_kde(dict_models_measures)
x_real, x_sepr, x_depr, x_ditras = [dict_x[m] for m in dict_models_measures.keys()]
y_real, y_sepr, y_depr, y_ditras = [dict_y[m] for m in dict_models_measures.keys()]
print('\n---------------------------------------- MEASURES DISTRIBUTIONS PLOT ----------------------------------------\n')
fig, ax = plt.subplots(nrows=2, ncols=3, figsize=(18, 9))
k = 0
for i in range(2):
for j in range(3):
# Calculating RMSE for each possible pair of the distributions
rmse_sepr_depr = round(mean_squared_error(y_sepr[k], y_depr[k], squared = False),4)
rmse_sepr_ditras = round(mean_squared_error(y_sepr[k], y_ditras[k], squared = False),4)
rmse_depr_ditras = round(mean_squared_error(y_depr[k], y_ditras[k], squared = False),4)
rmse_real_sepr = round(mean_squared_error(y_real[k], y_sepr[k], squared = False),4)
rmse_real_depr = round(mean_squared_error(y_real[k], y_depr[k], squared = False),4)
rmse_real_ditras = round(mean_squared_error(y_real[k], y_ditras[k], squared = False),4)
ax[i,j].plot(x_sepr[k], y_sepr[k], marker = 's', linestyle = 'dotted', linewidth = 2, markersize = 5, label = 'S-EPR')
ax[i,j].plot(x_depr[k], y_depr[k], marker = 'o', linestyle = 'dotted', linewidth = 2, markersize = 5, label = 'D-EPR')
ax[i,j].plot(x_ditras[k], y_ditras[k], marker = '^', linestyle = 'dotted', linewidth = 2, markersize = 5, label = 'Ditras')
ax[i,j].loglog()
ax[i,j].set(title = list_measures[k])
ax[i,j].set(ylabel = "$P({})$".format(list_labels[k]))
ax[i,j].set(xlabel = "${}$".format(list_labels[k]))
# Style for text box showing RMSE inside plot
props = dict(boxstyle = 'round', facecolor = 'wheat', alpha = 0.3)
# Showing the RMSE results, depending if the real measures are taken into account
if show_real == True:
ax[i,j].plot(x_real[k], y_real[k], marker = 'D', linestyle = 'dotted', linewidth = 2, markersize = 5, color = 'black', label = 'Real')
ax[i,j].text(-.6, .3, fontsize = 12, transform = ax[i,j].transAxes, bbox = props, s = ('RMSE M{} ${}$\n'
'\n'
'R/S:{}\n'
'R/D:{}\n'
'R/Di:{}\n'
'\n'
'S/D:{}\n'
'S/Di:{}\n'
'D/Di:{}').format(str(k+1),list_labels[k],
rmse_real_sepr,
rmse_real_depr,
rmse_real_ditras,
rmse_sepr_depr,
rmse_sepr_ditras,
rmse_depr_ditras))
else:
ax[i,j].text(-.6, .4, fontsize = 12, transform = ax[i,j].transAxes, bbox = props, s = ('RMSE M{} ${}$\n''\nS/D:{}\n''S/Di:{}\n''D/Di:{}').format(str(k+1),
list_labels[k],
rmse_sepr_depr,
rmse_sepr_ditras,
rmse_depr_ditras))
ax[i,j].legend()
fig.tight_layout()
k += 1
